Mary Tai Knox
130 Hudson Hall
Box 90291
k n o x m AT icsi DOT berkeley DOT edu

Office Location: FCIEMAS 3453

I am a research scientist in the SSPACISS laboratory at Duke University. Prior to Duke, I was a graduate student in electrical engineering at UC Berkeley, and worked at the International Computer Science Institute (ICSI). My PhD advisors were Nikki Mirghafori, Gerald Friedland, and Nelson Morgan.

conference papers
  • M. Knox, N. Mirghafori, and G. Friedland, "Where did I go wrong?: Identifying toublesome segments for speaker diarization systems," in Interspeech, 2012. pdf
  • M. Knox and G. Friedland, "Multimodal Speaker Diarization Using Oriented Optical Flow Histograms," in Interspeech, 2010. pdf
  • M. Knox, N. Morgan, and N. Mirghafori, "Getting the Last Laugh: Automatic Laughter Segmentation in Meetings," in Interspeech, 2008. pdf
  • M. Knox and N. Mirghafori, "Automatic laughter detection using neural networks," in Interspeech, 2007. pdf
  • M. Knox, "Automatic laughter segmentation," Master's thesis, UC Berkeley, 2008. pdf
  • EE40/43 (Circuits): Fall 2005, Summer 2006
  • EE20 (Intro to Signals and Systems): Spring 2006, Fall 2006, Spring 2007
  • EE120 (Signals and Systems): Fall 2007
Spring 2009
  • STAT244: Statistical Computing
Fall 2008
  • CS280: Computer Vision
    Course Project: Using Weighted Oriented Optical Flow Histograms for Multimodal Speaker Diarization pdf
Fall 2007
  • CS281A: Statistical Learning Theory
    Course Project: Automatic Laughter Segmentation pdf
Spring 2007
  • EE225D: Audio Signal Processing
    Course Project: Improving Frame Based Automatic Laughter Detection pdf
Fall 2006
  • CS294: Practical Machine Learning
    Course Project: Automatic Laughter Detection pdf
  • MUS208A: Music Perception and Cognition
  • Course Project: Automatic Laughter Detection pdf
Spring 2006
  • EE225A: Digital Signal Processing
    Course Project: Nonlinear and Non-Gaussian Estimation with a focus on Particle Filters slides
  • EE225B: Digital Image Processing
    Course Project: Enhanced Qslim for 3D City Model Compression and Rendering pdf
  • EE298: DSP/COM Seminar
Fall 2005
  • EE226A: Random Processes
  • EE221A: Linear Systems
  • EE301: Teaching Techniques